A Monosemantic Attribution Framework for Stable Interpretability in Clinical Neuroscience Large Language Models

2026-01-25

AI Summary

This work addresses the instability and inconsistency of attribution-based explanations in large language models (LLMs) within clinical neuroscience, which often arise from representational polysemy. To mitigate this issue, the authors propose a unified framework that integrates attribution with mechanistic interpretability by constructing and optimizing a monosemantic embedding space at specific LLM layers. This approach explicitly disentangles semantic features, thereby reducing variability across different attribution methods and yielding stable importance scores aligned with input–output mappings. By combining monosemantic representation learning, attribution ensembling, and layer-wise feature optimization, the method significantly enhances the reliability and reproducibility of explanations in tasks such as Alzheimer’s disease progression diagnosis, offering a trustworthy foundation for the safe deployment of LLMs in cognitive health applications.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: (Large) Language Models
• Computer Vision: Interpretability, Explainability, and Transparency

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Interpretability remains a key challenge for deploying large language models (LLMs) in clinical settings such as Alzheimer's disease progression diagnosis, where early and trustworthy predictions are essential. Existing attribution methods exhibit high inter-method variability and unstable explanations due to the polysemantic nature of LLM representations, while mechanistic interpretability approaches lack direct alignment with model inputs and outputs and do not provide explicit importance scores. We introduce a unified interpretability framework that integrates attributional and mechanistic perspectives through monosemantic feature extraction. By constructing a monosemantic embedding space at the level of an LLM layer and optimizing the framework to explicitly reduce inter-method variability, our approach produces stable input-level importance scores and highlights salient features via a decompressed representation of the layer of interest, advancing the safe and trustworthy application of LLMs in cognitive health and neurodegenerative disease.

Problem

• interpretability
• large language models
• attribution
• clinical neuroscience
• Alzheimer's disease

Innovation

• monosemantic
• attribution
• interpretability
• large language models
• clinical neuroscience